ABSTRACT
Chagas disease is caused by the protozoan parasite Trypanosoma cruzi. Assessment of parasitological cure upon treatment with available drugs relies on achieving consistent negative results in conventional parasitological and serological tests, which may take years to assess. Here, we evaluated the use of a recombinant T. cruzi antigen termed trypomastigote small surface antigen (TSSA) as an early serological marker of drug efficacy in T. cruzi-infected children. A cohort of 78 pediatric patients born to T. cruzi-infected mothers was included in this study. Only 39 of the children were infected with T. cruzi, and they were immediately treated with trypanocidal drugs. Serological responses against TSSA were evaluated in infected and noninfected populations during the follow-up period using an in-house enzyme-linked immunosorbent assay (ELISA) and compared to conventional serological methods. Anti-TSSA antibody titers decreased significantly faster than anti-whole parasite antibodies detected by conventional serology both in T. cruzi-infected patients undergoing effective treatment and in those not infected. The differential kinetics allowed a significant reduction in the required follow-up periods to evaluate therapeutic responses or to rule out maternal-fetal transmission. Finally, we present the case of a congenitally infected patient with an atypical course in whom TSSA provided an early marker for T. cruzi infection. In conclusion, we showed that TSSA was efficacious both for rapid assessment of treatment efficiency and for early negative diagnosis in infants at risk of congenital T. cruzi infection. Based upon these findings we propose the inclusion of TSSA for refining the posttherapeutic cure criterion and other diagnostic needs in pediatric Chagas disease.
INTRODUCTION
Chagas disease, caused by the protozoan Trypanosoma cruzi, is a lifelong disease for which no vaccines are yet available. With ∼6 million people already infected and up to 70 million individuals at risk of infection, Chagas disease constitutes one of the most important parasitic diseases in Latin America and an emerging threat to global public health (1, 2). T. cruzi transmission occurs when humans are exposed to the contaminated feces of blood-sucking triatomine vectors, through the ingestion of tainted food/beverages (3), blood transfusion or organ transplantation (1) or maternal-fetal transmission (4, 5). According to epidemiological data, maternal-fetal transmission occurs in ∼5% of T. cruzi-infected mothers, which leads to ∼15,000 new congenital cases per year (5).
Only two trypanocidal drugs, benznidazole (BZ) and nifurtimox (NX), are currently available for chemotherapy. Both are oral compounds that may display adverse effects (e.g., allergic dermatitis) and that cannot be used to treat pregnant women due to their uncertain teratogenic risks (5). Most importantly, BZ and NX show high efficacy only if administered at the onset of infection (6). The sole accepted criterion of cure relies on consistent negative results using conventional parasitological and serological tests (2). However, a significant proportion of patients have negative parasitological tests prior to treatment, thus making a subsequent negative result uninformative. PCR-based methods were proven useful in certain clinical situations usually associated with patent blood parasitemia such as congenital infections or disease reactivation in immunosuppressed patients (7–10). However, these methods remain to be clinically validated and are not yet available in regular health care centers. Moreover, some apparent false-positive results due to transplacental transfer of maternal parasite DNA have been described (7).
Conventional serological techniques such as enzyme-linked immunosorbent assay (ELISA) that use crude parasite homogenates are routinely used to assess posttherapeutic responses. However, seronegativization may take months to years to assess, even in successful treatments (11). Conventional serology methods display low predictive value for diagnosis and/or treatment evaluation of congenital infections until 8 to 9 months after birth due to the passive transfer of maternal antibodies (4, 5). Aiming to develop reliable posttherapeutic markers, different biochemical and serological approaches have been explored (12–27). The latter included the evaluation of T. cruzi antigenic fractions or defined antigens that elicit serological responses with different qualitative, quantitative, and/or kinetic properties. Overall, the best results were obtained with the F2/3 fraction (28, 29), which consists of highly antigenic α-galactosyl epitopes from the surface coat of bloodstream trypomastigotes (30). Several methodological drawbacks (i.e., costly and difficult purification procedures), however, preclude its routine implementation in clinical settings.
In previous works, we characterized a surface adhesion molecule from T. cruzi bloodstream trypomastigotes termed trypomastigote small surface antigen (TSSA) (31–33). TSSA elicits a strong humoral response during human infections (31, 34–36) and has been validated for Chagas disease serodiagnosis (37). At variance with F2/3, most anti-TSSA antibodies are directed to peptide epitopes (33, 36, 38), thus enabling the straightforward production of a highly pure diagnostic reagent in engineered bacteria. Here, we evaluated the potential use of recombinant TSSA as a novel serological marker of drug efficacy in T. cruzi-infected children.
RESULTS
A total of 78 children (4 days to 10 years old) born to T. cruzi-infected mothers were included in this study. Thirty-eight of them were initially diagnosed as infected with T. cruzi and were coursing either the acute or the early chronic phase of Chagas disease, with no evidence of cardiac abnormalities or any other Chagas disease-associated pathology. These 38 T. cruzi-infected patients were split into 2 groups based on their age range at diagnosis. Group 1 comprised 26 T. cruzi-infected children (0.59 to 9 years old; median, 4.5 years old) that were diagnosed by conventional serology whereas group 2 comprised 12 T. cruzi-infected babies (8 to 143 days old; median, 36 days old) that were diagnosed by parasitological tests. A total of 430 serum samples were obtained from these patients during treatment/follow-up (median, 12 samples per patient). Samples were analyzed by conventional serology and, whenever possible, by PCR. The average follow-up time (and range) for these patients was 36 months (14.57 to 111.53 months). Chemotherapy was considered successful in every T. cruzi-infected patient, based on a steady decrease in conventional serology values during the follow-up period and, in most cases (24/26 from group 1 and 5/12 from group 2), based also on PCR negativization.
Group 3 included 40 infants (4 to 118 days old; median, 31 days old) born to T. cruzi-infected mothers. At variance with group 2 patients, these patients rendered negative results for parasitological tests. A total of 148 samples were analyzed by conventional serology during the follow-up (median, 4 samples per follow-up). Conventional serology became negative in group 3 infants at the end of the follow-up period except for patient REC52, who was accordingly excluded from this group. A flow chart summarizing this information is depicted in Fig. 1; and all relevant demographic, clinical, and diagnostic features of every patient included in this study, and of their mothers (when available), are summarized in Tables S1 to S4 in the supplemental material.
Study population, inclusion criteria, and group composition.
Recombinant TSSA-based ELISA for assessing therapy efficacy in T. cruzi-infected children.Serological reactivity toward TSSA in T. cruzi-infected children was first assessed in samples taken at diagnosis (pretreatment). Nineteen of 26 (73%) and 10 of 12 (83%) children belonging to groups 1 and 2, respectively, yielded positive results (Table S5) and were hence evaluated for anti-TSSA antibody titers during the serologic follow-up. Recombinant TSSA-based ELISA (TSSA-ELISA) and total parasite homogenate-based ELISA (tELISA) results are shown in Tables S1 and S2 and Fig. S1; linear regression analyses of these data are shown in Fig. 2. Overall, patients from group 1 showed a steady decrease in anti-T. cruzi antibody titers after treatment, which in certain cases led to seronegativization. This decreasing trend was not significantly different when assessed by tELISA or TSSA-ELISA (P = 0.28) (Fig. 2A). However, upon stratification of group 1 by age and hence by duration of infection, significant differences in the serological regression slopes for either method were detected for the younger patients (1 to 4 years old; P = 0.01) but not the older ones (4 to 10 years old; P = 0.6) (Fig. 2B and C). Patients from group 2 displayed significant differences (P = 0.01) in the serological regression slopes assessed by TSSA-ELISA or tELISA (Fig. 2D).
Serological regression analysis of all patients from group 1 (A), younger patients (0.59 to 4 years old) from group 1 (B), older patients (4 to 10 years old) from group 1 (C), and patients from group 2 (D). tELISA and TSSA-ELISA results (expressed as % of the first sample) are indicated in solid and dashed lines, respectively. Mean reactivity and SD values for each time point are shown in solid (tELISA) and open (TSSA-ELISA) circles. Slope (95% CI) and R2 values are indicated for each data set. P, pretreatment. ANCOVA analyses were performed to compare slopes.
A total of 22 T. cruzi-infected patients (10/26 from group 1 and 12/12 from group 2) seronegativized following treatment as shown by tELISA. Of these, only 18 (8 from group 1 and 10 from group 2) were TSSA reactive (Table S5). Interestingly, seronegativization for these 18 patients occurred either before (n = 16) or at the same time (n = 2) in TSSA-ELISA testing as in tELISA testing. Moreover, 3 patients from group 1 achieved seronegativization in TSSA-ELISA but not in tELISA (denoted as censored cases in Fig. 3A). Kaplan-Meier curves comparing the performance of both methods among seronegativized patients are plotted in Fig. 3. As shown, the median time values of negativization for TSSA-ELISA and tELISA were 8.67 and 32 months, respectively, for group 1 (P < 0.0001); and 2.21 and 5.4 months, respectively, for group 2 (P = 0.002). Again, significant differences in the median time values of negativization for either method were detected for the younger patients (P<.0001) but not for the older ones (P = 0.2) upon stratification of group 1 (Fig. 3B).
Kaplan-Meier curves of seronegativized patients from group 1 (A) and group 2 (C). Upon stratification of group 1 by age, similar analysis was performed for younger (0.59 to 4 years old; violet) and older (4 to 10 years old; orange) patients (B). tELISA and TSSA-ELISA results are indicated in solid and dashed lines, respectively. Median (95% CI) values are indicated for each data set. Censored cases are indicated with square symbols. Log-rank (Mantel-Cox) analyses were performed to compare median time of seronegativization. N/A, confidence interval was not calculated due to the small number of samples.
Comparative analysis of tELISA data indicated that serological regression followed distinct kinetics, being significantly faster (P < 0.0001) in group 2 (slope −10.53) than in group 1 (slope −1.527) (Fig. 2). This in turn translated into significantly shorter (P < 0.0001) time periods to reach the tELISA negativity threshold (median values, 5.4 months and 32 months for groups 2 and 1, respectively) (Fig. 3). Along the same lines, TSSA-ELISA revealed differences in serological regression slopes for group 2 (n = 10) compared to those for group 1 (n = 19) (−20.62 and −2.058, respectively; P = 0.07) (Fig. 2) as well as shorter time periods to achieve seronegativization (2.21 months and 8.67 months, respectively; P = 0.0001) (Fig. 3). Overall, the latter results support previous findings indicating that the decline in anti-T. cruzi antibody titers after chemotherapy is faster in younger children (6, 39).
TSSA-ELISA for early assessment of congenital T. cruzi transmission.According to current guidelines, ruling out maternal-to-fetal T. cruzi transmission requires negative results in parasitological tests performed early after birth and in conventional serologic tests carried out at 8 to 9 months of life, upon clearing of antibodies of maternal origin (5). To explore whether TSSA might also improve diagnosis in this area, data from the serologic follow-up of patients from group 3 were analyzed as before. TSSA-ELISA (n = 36 since 3 patients were born to TSSA-non-reactive mothers [Table S5]) and tELISA (n = 39) results are shown in Table S3 and Fig. S1; linear regression analyses of these data are plotted in Fig. 4. Both kinds of maternally transferred antibodies showed a steady decline early after birth, although with significant differences in their regression slopes (Fig. 4A). Accordingly, patients from group 3 seronegativized either before (n = 31) or at the same time (n = 5) in TSSA-ELISA testing as in tELISA testing, and they displayed significantly different median values of seronegativization (Fig. 4B). Interestingly, patients from groups 2 and 3 displayed almost indistinguishable median values of seronegativization assessed by either tELISA (P = 0.34) or TSSA-ELISA (P = 0.45) (Fig. 4C), suggesting that, if treated immediately after birth, T. cruzi-infected children do not elicit robust, parasite-specific serological responses. In such a scenario, the kinetics of seronegativization seem to be mainly driven by the persistence of passively transferred maternal IgG antibodies.
Serological regression analysis (A) and Kaplan-Meier curves (B) comparing seronegativization of patients from group 3 determined by tELISA (solid lines) with that determined by TSSA-ELISA (dashed lines). (C) Seronegativization of group 2 patients (green lines) compared to that of group 3 patients (blue lines) as determined by either tELISA (solid lines) or TSSA-ELISA (dashed lines). Mean reactivity and SD values for each time point are shown in solid (tELISA) and open (TSSA-ELISA) circles. Slope (95% CI) and R2 values (A) or median (95% CI) values (B) are indicated for each data set. ANCOVA and log-rank (Mantel-Cox) analyses were performed to compare slopes and median time of seronegativization, respectively.
One of the patients originally assigned to group 3 yielded particular results which deserve to be analyzed separately. This patient, labeled REC52 (Fig. 1), was born and raised within the urban limits of Buenos Aires, an area free of vector-borne parasite transmission, and did not undergo blood transfusion. Despite being positive when tested by PCR-based methods, REC52 displayed consistent negative results in conventional parasitological tests carried out early after birth (Fig. 5 and Table S4). At 10.5 months of age, and having achieved seronegativization for conventional serological methods, REC52 was declared noninfected and the follow-up was ended. Seronegativization was achieved at 4.2 and 7.2 months of life as measured by TSSA-ELISA and tELISA, respectively, values that are well within the range of noninfected children included in group 3 (Fig. 4). Unexpectedly, however, REC52 yielded positive results for TSSA-ELISA at 10.5 months of age (Fig. 5). At this time point, we also detected reactivity toward shed acute-phase antigen (SAPA) (40), the canonical acute-phase T. cruzi antigen (Fig. 5). Positive results for TSSA-ELISA were confirmed at 19 months of age (Fig. 5). At this time, REC52 also yielded positive results for tELISA, and treatment with BZ was thereby initiated. Following treatment, REC52 showed typical anti-T. cruzi antibody decay, indicating therapeutic efficacy (Fig. 5).
TSSA-ELISA (black circles), tELISA (gray triangles), SAPA-ELISA (gray squares), IHA, and PCR results for patient REC52 are shown. Arrow, treatment initiation; dotted line, cutoff determined for both TSSA-ELISA and SAPA-ELISA; +, positive; −, negative. N/D, not done.
DISCUSSION
Identification of novel and reliable posttherapeutic markers is an urgent need in the field of Chagas disease (5, 28, 29). As shown here, TSSA-ELISA provides a significantly better indicator of trypanocidal drug efficacy than currently used serological methods, particularly for newborns and infants (Fig. 2 and 3). Unfortunately, TSSA results are difficult to compare to those reported for other T. cruzi recombinant antigens and/or antigenic fractions, since most of these studies did not involve newborns and infants but older T. cruzi-infected populations (12–21). Nevertheless, compared to the only similarly designed study that we are aware of, the median times of seronegativization for TSSA (2.21 and 8.67 months for children under and over 8 months of age, respectively) were significantly lower than those recorded for F2/3 (4 and 21.9 months, respectively) (41). As mentioned, F2/3 is so far considered the best alternative serological marker for treatment evaluation in Chagas disease (16, 28, 29).
We show that in addition to providing a novel tool able to shorten follow-up periods following chemotherapy, TSSA improves diagnosis in infants at risk of congenital T. cruzi infection. Moreover, our findings with REC52, although preliminary, suggest the applicability of TSSA as an alternative early marker for T. cruzi infection in certain clinical situations. Based on PCR results and clinical history, we postulate that REC52 became congenitally infected with T. cruzi, although the course of this infection was subclinical and below the detection limits of different parasitological and serological tests until “reemergence,” probably due to the disappearance of maternal antibodies. As shown in Fig. 5, TSSA-ELISA also displayed better performance than conventional serology for the detection of this infection reemergence.
One issue that needs to be addressed in order to improve the clinical value of TSSA is that of its suboptimal sensitivity, which may be attributed to variations in the clinical, immunological, and/or immunogenetic features of patients, and/or to differences in the antigenic constitution of the infecting T. cruzi strain(s) (42). Structural differences among protein variants carried by distinct parasite strains have a major impact on TSSA antigenicity (31, 37, 43). In our study, for instance, TSSA prevalence was 86%, which is consistent with previous data (∼86 to 91%) (36, 37). In the case of the two TSSA-nonreactive children from group 2, however, it is most notable that they were born to TSSA-reactive mothers (Table S2). It may be therefore hypothesized that both of them underwent clearing of anti-TSSA antibodies of maternal origin at some point between birth and initial T. cruzi infection diagnosis. In such a case, the sensitivity and overall performance of TSSA-ELISA would have been underestimated. Despite these considerations, different alternatives, including the use of a mixture of TSSA variants, are currently being explored to improve the clinical value of TSSA-ELISA.
Progress toward development of novel and better treatments for Chagas disease has been slow and usually disappointing (44, 45). This scenario fortunately seems destined to change in the coming years with the recent development of robust tools to screen, prioritize, and evaluate novel antitrypanosomal drugs (46–48). Identification of biomarkers able to refine the posttherapeutic criterion is instrumental to hasten the assessment of current trypanocidal chemotherapies and, most importantly, for the development of much needed novel and improved treatments.
MATERIALS AND METHODS
Study population and screening for T. cruzi infection.A cohort of 78 children of both sexes and born to T. cruzi-infected mothers was recruited for this study. All of them were screened for T. cruzi infection and followed up at the Servicio de Parasitología-Chagas, Hospital de Niños Dr. Ricardo Gutierrez following current norms. Briefly, T. cruzi infection in children over 8 months of age was diagnosed using two conventional serological tests: an ELISA that uses crude parasite homogenates (tELISA) (Wiener Chagatest-ELISA) and an indirect hemagglutination assay (IHA) (Wiener Chagatest-HAI). Both are validated commercial tests widely used in clinical settings. Infection in children under 8 months of age was assessed by the microhematocrit method (4). In case of positive results, patients were treated immediately. In case of negative parasitological results, children were called for a medical appointments at 3, 6, and 9 months of age. Serum samples were taken at each time point and analyzed by conventional serological tests. Those patients displaying negative results for conventional serology at 9 months were considered noninfected whereas those displaying positive results were immediately treated. Most of the participating children were born and raised within the urban limits of Buenos Aires, Argentina, an area free of vector-borne parasite transmission, and hence most likely acquired T. cruzi infection congenitally.
Treatment and follow-up. T. cruzi-infected children were treated with BZ (5 to 8 mg/kg twice a day [b.i.d]) or NX (10 to 15 mg/kg three times a day [t.i.d]) (49). Infants' doses were provided as fractioned tablets (BZ [Abarax] as 100-mg tablets, Elea, Argentina, or NX [Lampit] as 120-mg tablets, Bayer) and treatment was open label for 60 days. Medication was provided in monthly batches, and compliance was assessed by tablet counting at each visit. Caregivers were also provided with a treatment diary to record doses administered, times of doses, symptoms, and problems associated with the treatment. Serum samples were taken at diagnosis (pretreatment); at 7, 30, and 60 days (during treatment); and every 3 to 6 months after treatment (follow-up). A detailed clinical history, physical examination, and routine laboratory tests were conducted during treatment (49), and T. cruzi conventional serology was carried out at every medical visit during the follow-up period. DNA was purified from whole blood samples and used as the template for a multiplex real-time PCR targeting a 166-bp segment from T. cruzi satellite DNA as described previously (50).
Recombinant TSSA-based ELISA (TSSA-ELISA).The glutathione S-transferase (GST)-fusion protein bearing the antigenic region (residues 24 to 62) of T. cruzi (CL Brener clone) TSSA has been described (36). GST-TSSA was expressed in Escherichia coli and purified from the soluble fraction to almost homogeneity by a single glutathione affinity chromatography step (36). Flat-bottomed 96-well Nunc-Immuno plates (Nunc, Roskilde, Denmark) were coated overnight at 4°C with 80 μl of GST-TSSA dissolved in carbonate buffer (pH 9.6) at 0.25 μg/ml and processed for a previously validated, colorimetric TSSA-ELISA as described previously (36). Serum samples were assayed in duplicate at 1:500 dilution, and those displaying (mean −3 SD) a value greater than the corresponding cutoff value (calculated as the [mean + 3 SD] of 4 samples from healthy children born to non-Chagasic mothers) were considered reactive. Reactivity of samples used to determine the cutoff ranged from 0.06 to 0.12 absorbance units (36). For comparison purposes, cutoff and sample values were expressed as a percentage of a positive control (a chronic Chagasic patient yielding 0.8 to 1.4 absorbance units) included in each assay (36). The overall performance of our TSSA-ELISA has been extensively validated (37). When indicated, anti-SAPA (40, 51) IgG responses were evaluated by an in-house ELISA (36).
Statistical treatment of results.A linear regression model was used to examine the course of antibody levels over time. Because tELISA and TSSA-ELISA optical density (OD) values were highly variable among different patients, even among those from the same group, they were expressed as a percentage of the OD value of the first specimen, which was a pretreatment sample (for patients from groups 1 and 2) or the corresponding mother sample (for patients from group 3). The reactivity of negative samples was expressed as zero. For each group of patients and each method, a slope parameter with 95% confidence interval (CI) was calculated based upon time point data for which at least one patient per group was positive. In instances when two or more consecutive samples were nonreactive by either tELISA or TSSA-ELISA, the date of the first negative sample was considered the time of seronegativization for this method. Kaplan-Meier curves and linear regression analysis were plotted and compared using the log-rank (Mantel-Cox) test to obtain median time of seronegativization or analysis of covariance (ANCOVA), respectively; both available in GraphPad Prism 5 software version 5.01 for Windows (San Diego, CA). CIs were calculated using SPSS Statistics version 20 (IBM).
Ethics statement.The study protocol was approved by the research and teaching committee and the bioethics committee from the Hospital de Niños Dr. Ricardo Gutierrez. Written informed consent was required from each patient's legal representatives as well as assent from the patient, if applicable. All samples were decoded and deidentified before they were provided for research purposes.
ACKNOWLEDGMENTS
This investigation received financial support from the Agencia Nacional de Promoción Científica y Tecnológica, Argentina (www.agencia.mincyt.gob.ar ) to C.A.B. and J.A. and the Fundación Bunge y Born, Argentina (www.fundacionbyb.org ) to C.A.B. V.B. and L.J.M. hold fellowships from the Argentinean Research Council (CONICET), and M.B., A.E.C., C.A.B., and J.A. are career investigators from the same institution. In addition, J.A. is a Clinical Investigator from the Hospital de Niños Dr. Ricardo Gutiérrez (Ministerio de Salud, Gobierno de la Ciudad de Buenos Aires).
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
FOOTNOTES
- Received 16 August 2017.
- Returned for modification 11 September 2017.
- Accepted 27 September 2017.
- Accepted manuscript posted online 4 October 2017.
Supplemental material for this article may be found at https://doi.org/10.1128/JCM.01317-17 .
- Copyright © 2017 American Society for Microbiology.
